University of Glasgow's GenAI assessment guidance sharpens what student feedback should evidence
Updated Jul 23, 2026
University of Glasgow's GenAI assessment guidance matters because it turns broad AI principles into a model students will actually meet at assessment level. On 7 July 2026, the university published GenAI for Assessment Guidance in Academic Year 2026-27, setting out how its updated rules will apply from the start of the 2026-27 academic year. For teams responsible for student voice, the practical implication is direct: if universities want better evidence on whether AI rules feel fair, clear, and consistent, they now need feedback routes that test those questions explicitly.
What has changed in Glasgow's GenAI assessment guidance
The new GenAI assessment guidance creates a default position for two different assessment contexts. Glasgow says supervised assessment, such as invigilated exams, in-class tests, practical work, and vivas, does not permit GenAI unless staff expressly allow it. Unsupervised assessment, such as essays, reports, projects, portfolios, and dissertations, generally permits GenAI, but students must acknowledge its use and course coordinators can still set tighter limits or prohibit it for a specific task. That matters because the university is not relying on one generic AI statement. It is tying the rules to assessment format and then allowing module-level exceptions where needed.
"The assessment guidance provided should clearly indicate which of the two scenarios the assessment falls into."
The operational detail is just as important. Glasgow says staff should state the scenario and any specific conditions in advance of students starting both formative and summative work, and that updated student guidance will be shared centrally ahead of the new academic year through Moodle, assessment briefs, and in-class discussion. That shifts the burden away from students trying to infer what a vague institutional AI policy means in practice. It also means course teams now have a clearer obligation to communicate the rules at assessment level, not only at school or university level.
The wider institutional position is also worth noting. On its student AI guidance page, Glasgow says students will graduate into an AI-augmented world and should learn to use these tools ethically, critically, and transparently rather than simply being told to avoid them. This is not a prohibition model. It is a structured use model. For universities across the sector, that makes the story more useful than a generic AI update. It shows one institution trying to balance academic integrity, assessment design, and student clarity through a model that is specific enough to test.
What this means for institutions
The first implication is that consistency now becomes an implementation problem, not just a policy problem. If course coordinators can still vary the rules assessment by assessment, then universities need to know where those exceptions sit, how they are phrased, and whether students understand them. That is closely aligned with the recent QAA warning about AI as a student experience risk: inconsistency often appears inside programmes and modules before it becomes visible in a policy review or a complaints trend.
The second implication is that student feedback collection needs to get more precise. Institutions should not ask only whether students feel positive or negative about AI. They need to know whether students understood which scenario applied, whether the rules were explained early enough, whether acknowledgement expectations were workable, and whether staff explanations matched the assessment brief. A student comment analysis governance checklist is useful here because it forces teams to define which routes collect that evidence, who reviews it, and how conflicting signals are escalated.
The third implication is timing. Glasgow is implementing this model from the start of 2026-27, which means the first live evidence will arrive quickly through induction feedback, module evaluations, rep channels, and early assessment questions. Universities making similar changes should treat those early comments as diagnostic evidence, not as noise. If students say one module allows AI brainstorming, another expects full disclosure, and a third says nothing at all, that is not a minor drafting issue. It is a sign that the guidance may be coherent on paper but uneven in delivery.
How student feedback analysis connects
This is exactly the kind of change where open-text feedback becomes more useful than headline scores. Students will describe where briefs stayed vague, where staff guidance changed between classes, what counted as acceptable acknowledgement, and whether the supervised versus unsupervised distinction made sense in practice. That is why Glasgow's announcement sits naturally alongside the recent sector push for clearer student evidence on AI-era assessment and feedback. If assessment rules are changing quickly, institutions need to know not just whether students noticed, but where the new model still feels unclear or uneven.
For teams reviewing those comments at scale, the method matters as much as the theme. If the institutional risk is inconsistency, the analysis workflow should not introduce more of it. Student Voice Analytics is one practical way to keep that work reproducible across large comment sets, but the broader point is methodological: universities need student feedback evidence they can revisit, compare, and use to tighten assessment guidance before confusion turns into mistrust.
FAQ
Q: What should institutions do now?
A: Start by mapping which assessments would fall into each scenario and where local exceptions are likely to arise. Then review assessment brief templates, acknowledgement guidance, and staff-facing instructions so students hear the same message in briefs, Moodle, and class. Finally, add a few targeted prompts to your existing student feedback routes so you can test whether the rules felt clear in practice.
Q: What is the timeline and scope of Glasgow's GenAI assessment guidance?
A: Glasgow updated the guidance in May 2026 and published its implementation note on 7 July 2026. The two-scenario model will apply from the start of the 2026-27 academic year. The immediate scope is one Scottish university, but the operational model is relevant across UK higher education because many institutions are trying to move from broad AI principles to module-level assessment rules.
Q: What is the broader implication for student voice?
A: Student voice is likely to become one of the main ways institutions test whether AI guidance is actually legible at assessment level. A university can publish a clear central policy and still create confusion if module briefs, staff explanations, and acknowledgement expectations drift apart. Feedback evidence is what shows where that drift is happening.
References
[University of Glasgow]: "GenAI for Assessment Guidance in Academic Year 2026-27" Published: 2026-07-07
[University of Glasgow]: "GenAI Guidance 2026: Assessment Scenarios" Published: not stated
[University of Glasgow]: "Artificial Intelligence" Published: not stated
Request a walkthrough
Book a free Student Voice Analytics demo
See all-comment coverage, sector benchmarks, and reporting designed for OfS quality and NSS requirements.
- All-comment coverage with HE-tuned taxonomy and sentiment.
- Versioned outputs with TEF-ready reporting.
- Benchmarks and BI-ready exports for boards and Senate.
UK-hosted · No public LLM APIs · Same-day turnaround